Tuesday, January 17, 2012

9 Ways Journalists Show They Don't Get Science

Over on Guardian blogs, there's hot discussion about perceptions of hacks and lab geeks, who often seem to work at cross purposes. It was sparked by a post authored by the online editor of Nature, Ananyo Bhattacharya.   The comment below turned the discussion around and drew widesread approval from academics.  Definitely there are some salient points.  Hat tip to "jferdy5" for his/her insights on science reporters' shortcomings, listed below:
  • 1. False equivocation: one of the reasons why there's doubt about global warming is because journalists take a handful of crackpots, many of which have never studied science, and then "equivocates," or gives roughly equal weight, to this argument versus broad scientific concensus.
  • 2. Anecdotes do not equal systematic evidence (they're inferior): "while vaccinations may work for millions of people, that will not help little Billy who now has autism." Sounds familiar?
  •  3. Manufacturing dissent: taking a fringe opinion (anti-Global warming, anti-vaccination, anti-evolution, etc) and giving it a disproportional amount of space. 
  •  4. Expert opinion: articles should clearly define why Dr. X is considered an expert. Frequently, I find they're either an economist, statistician, or other person who hasn't published in a particular area and seems to be motivated by political views rather than dispassionate scientific debate.
  •  5. Data journalism: this is a particularly insipid form of pseudo-science: journalists should not simply tabulate national rates based on their own analyses because they do not do the correct statistical tests (Fisher exact tests, etc) to check if a result is scientifically valid or not. It's bad to do this because the result may look scientifically valid when it is not.
  •  6. The use of "narrative:" linking together anecdotes using emotive language manipulates and misleads readers. In law it's called "leading the witness," in science, "systematic bias." In journalism, why is it considered "good writing?"
  •  7. No clear authors of articles: it's hard to examine the validity of science "reporting" if we do not know who is doing it. Simply putting "Guardian" or "New York Times" does not help. Scientists are required to disclose their names and sources of funding, journalists should do the same. 
  •  8. No citations: most Guardian articles, and in many papers, will discuss a study in inflammatory terms and not provide a link to the article, when they're available on pubmed. As well, no one cares what an untrained journalists' interpretation is. Just put the results and let us make up our mind. And no, sassy science headlines don't draw in readers, they just make your paper look stupid.
  •  9. Sending a reporter to a country does not trump public health statistics: if the WHO reports that malnutrition in Malawi or India is declining, it does not matter how many people your "Development" reporters interview, you cannot recalculate / dispute a national malnutrition / AIDS / TB rate based on a few subjective, systematically biased, personal interviews.
  •  The Guardian does all of these. Perhaps we can have a discussion about it, and why journalism is responsible for driving pseudo-science and the decline of Western civilization?

No comments: